clova-speech-lecture-mcp
CLOVA Speech Lecture MCP Server
Ein MCP-Server für STT (Speech-to-Text) und Zusammenfassungen von Vorlesungsaufzeichnungen.
Ruft die Naver CLOVA Speech / CLOVA Studio APIs über den Google ADK-Agenten auf.
Verfügbare Tools
Tool-Name | Beschreibung | Geeignet für |
| Synchrones STT (bis zu 60 Sek.) | Kurze Clips, Tests |
| Asynchrone STT-Einreichung (ganze Vorlesung) | Vorlesungen von mehreren Minuten bis Stunden |
| Abruf des asynchronen STT-Ergebnisses | Polling nach dem Submit |
| Zusammenfassung des Vorlesungstranskripts | Nach Abschluss der Transkription |
Related MCP server: Whisper Valet
Erste Schritte
1. API-Schlüssel abrufen
CLOVA Speech (STT)
Melden Sie sich bei der Naver Cloud Platform an.
AI Services → CLOVA Speech → Nutzung beantragen.
Nach der Diensterstellung den Secret Key kopieren.
CLOVA Studio (Zusammenfassung)
AI Services → CLOVA Studio → Nutzung beantragen.
App im Playground erstellen.
API-Schlüssel und App-ID kopieren.
2. .env-Konfiguration
cp .env .env.local # 실제 값으로 수정CLOVA_SPEECH_API_KEY=실제_키_입력
CLOVA_STUDIO_CLIENT_ID=실제_Client_ID
CLOVA_STUDIO_CLIENT_SECRET=실제_Client_Secret
CLOVA_STUDIO_APP_ID=실제_앱_ID3. Ausführung
# Docker
docker-compose up -d
# 로컬 실행
pip install -r requirements.txt
python app/main.py4. ADK-Agenten-Verbindung
# agent.py 예시
tools = [
MCPToolset(
connection_params=SseServerParams(url="http://localhost:8002/sse")
)
]Ablauf der Vorlesungsverarbeitung
[강의 오디오 파일]
↓
transcribe_lecture_submit(file_path, enable_diarization=True)
↓ (task_id 반환)
get_transcription_result(task_id) ← 완료까지 자동 폴링
↓ (full_text 반환)
summarize_lecture(text)
↓
[요약문 + 키워드]Unterstützte Audioformate
.wav .mp3 .flac .m4a .aac .ogg
Unterstützte Sprachen
Code (kurz) | Code (lang) | Sprache |
|
| Koreanisch |
|
| Englisch |
|
| Japanisch |
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
This server cannot be deployed
Maintenance
Related MCP Connectors
Pronunciation assessment, phoneme scoring, speaker voice ID, audio transcription, speech synthesis.
Pronunciation scoring, speech-to-text, and text-to-speech for language learning
Transcribe audio and video into speaker-labelled transcripts, subtitles, clips, and cited Q&A.
Transcribe, summarize and analyze audio with AssemblyAI speech-to-text and Speech Understanding.
Related MCP Servers
- AlicenseAqualityCmaintenanceProvides accurate meeting transcription with speaker diarization and multilingual support, allowing users to submit audio URLs, poll transcription status, get transcripts, and summarize via MCP tools in their IDE.81MIT
- AlicenseAqualityDmaintenanceEnables automated audio restoration, transcription, and speaker diarization via MCP tools for queuing files, monitoring progress, and retrieving speaker-labeled transcripts.9MIT

Tapra MCP serverofficial
AlicenseAqualityBmaintenanceEnables transcription and subtitling of Tunisian Derja mixed with French and English, speech and voiceover generation, translation, and Arabizi–Arabic script conversion inside MCP clients. It also supports speaker diarization, subtitle translation, transcript search, and audio editing.16488 npmMIT
AssemblyAI MCP Serverofficial
AlicenseNot gradedqualityBmaintenanceEnables MCP-capable agents to submit and retrieve audio transcriptions, apply speech understanding such as translation, speaker identification, summaries and action items, and enforce guardrails including PII redaction and content moderation.2MIT